US laboratory administers chemical treatments to enhance the energy capacity of electric vehicle batteries, prolonging their power duration
In a groundbreaking development, scientists at Argonne National Laboratory have utilised machine learning to identify new chemical "medicines" for batteries, specifically focusing on LNMO (lithium, nickel, manganese, and oxygen) batteries.
LNMO batteries, known for their higher energy capacity and absence of cobalt, a material with supply chain limitations, pose a unique challenge due to their operation at a high voltage of nearly 5 volts. This voltage is above the stability limit of most electrolytes, making them prone to decomposition.
To combat this issue, scientists employ electrolyte additives. These additives, when effective, decompose during the initial battery cycles to form a stable interface on the electrodes. By doing so, they lower resistance and reduce degradation, which can significantly improve the battery's performance.
The traditional belief is that a vast amount of data is necessary to train a machine learning model. However, the Argonne team's work demonstrates that a well-selected set of data, in this case, 28 additives, is sufficient. By training the model on this dataset, the system learned to recognise molecular features associated with certain battery metrics, such as resistance and energy capacity.
This innovative approach allowed the researchers to bypass an experimental process that would have taken four to six months. They were then able to predict the performance of 125 new chemical combinations using the model.
The Fraunhofer Institute for Silicate Research (Fraunhofer ISC) has also advanced the use of machine learning models to identify new chemical materials for batteries through automated and robot-assisted material development processes. This collaboration between institutions underscores the potential of machine learning in accelerating and enhancing battery research.
In conclusion, the development of this machine learning model marks a significant step forward in the field of battery research. It promises to expedite the discovery of new battery chemistries, potentially leading to more efficient, stable, and environmentally friendly energy storage solutions.
